TV, Violence and Aggression

By Mentor Cana on December 11, 2003 12:20 PM | Permalink | Comments (0) | TrackBacks (0)

In determining which of the four readings to analyze closer for this exercise, the Robinson, Wilde, Navracrus, Haydel and Varady (2001) article presents a more coherent research piece, in my viewpoint, primarily because the theoretical background is better understood (in comparison with the rest of the articles) considering that I’m not very well versed in behavioral and cognitive sciences which seem to be necessary to fully understand, appreciate and be able to provide constructive criticism. Further, Robinson et al. have gone to a great length to elaborate in details on their methodology, the measures used, and their rationale for using them, including a rather detailed report on the statistical procedure used with the corresponding results. The article ends with a great amount, relatively speaking, of concluding remarks including elaborations of limitations and strengths.

Unlike the other three articles that attempt to understand what happens with treatment group(s) when exposed to intervention that increases the dose of exposure to aggressive and violent media or exposure to media in general, the Robinson et al. article attempts to answer whether reduction in media exposure (reduced television, videotape and videogame use) has the effect to reduce violent and aggressive behaviors.

The basic premise in Robinson et al., as it has been shown by the rest of the articles, is that exposure to media increases violent and aggressive behavior (Centerwall, 1989), especially the exposure to violent and aggressive television and videotape viewing, results in the subjects to exhibit less sensitivity and concern about such behaviors when committed by others (Linz, Donnoerstein, & Penrod, 1984). Thus, Robinson et al. hypothesize that reduction in media exposure in general by reducing television, videotape and videogame use, reduces violent and aggressive behaviors in children.

The experiment design is overall very well thought. However, the experiment would have had more linking power (internal validity) if instead of one by one design, it was setup as two by two design. The way the experiment is set, it can’t be shown whether the resulting difference between the treatment group (one) and the control group (one) are due to the treatment or some other external or internal properties in the schools chosen. If there were two control and two treatment groups, it would have been possible to detect any potential changes of difference effects of the same treatment in two different schools.

Keeping the subjects within the same age group (or grade level) was a sound and perhaps necessary design decision in order to control for age difference. However, I think this was unnecessary restriction since if they had chosen two different age groups the results would have told us more about the effect of TV, videotape and videogame reduced exposure across different age groups. Obviously, this would have meant increasing the number of cases (subjects) in order to be able to achieve any significant statistics.

In terms of measurements, the design is also very well thought. The peer rating of aggressive behavior is important to bring forth the relative perception among the students. The measures of perception of mean and scary world are an attempt to understand the predisposition of students’ perception that might affect the outcome. Further, the parent measures, and the playground observations by eight trained observers for each subject are an attempt to triangulate the measure of aggressive behavior. It is interesting to note that the researchers were careful to establish inter-observer reliability by having eight (8) different observers observe and rate childrens’ aggressive behaviors.

However, I believe there is a weakness in the design in respect to the measures, at least from what has been reported in the paper, that there is no clear definition of what it means to observe an aggressive behavior: touching, about to touch with the intention to hit, actual hitting or about to hit, yelling or about to yell? Thus, what observers see as an aggressive behavior, might be different than the perception of such behaviors as perceived by the students and as perceived by their parents. In addition, the parent reporting is troublesome measure to rely on due to the potential bias by parents not reporting the aggressive behaviors of their children. This is further complicated by the fact that the level that divides the aggressive behavior from the non-aggressive one might be different in different families due to various personal and situational traits. To this extend, at treatment level, it is reported that the parents of the treatment school/group have college education 24% higher than the control group. It is quite possible that parents with higher college education might respond differently to what constitutes aggressive behavior. So, one of the drawbacks of this study is that it has not considered the implication of education and age. Then, it is hard to conclude whether the difference in aggressive behaviors at school from that at home is due to actual aggressive behavior difference or due to the difference that children might act more aggressive at school because they can do so.

Another concern with the Robinson’s et al. study is that the sample size is different for the different measures obtained, ranging from 104 cases for the playground observations (52+52) up to the total number of 225 cases (105+120) at pretest. An additional concern with the sample size is the ratio of variables collected to the sample size in each of the measures. There are at least 20 variables. This results in a low ratio of about 10 cases per variable.

The treatment delivery process has been also carefully designed with in school lectures about reducing TV, videotape, and videogame use delivered by their regular (but trained) classroom teacher. There is no report on the content of the lecturers. However, if the lectures were attempting to show the benefit of reduced TV, videotape and videogame use, it is quiet possible that the resulting effects were due to the media/information literacy received by the students rather than the turnoff period. It would have been beneficial to separate and assess the effects of the lectures separately from the effect of the turn-off period, and any possible interactions between the two.

In contrast to Robinson’s et al. study with intervention of staying away from TV, videotapes and videogames altogether, Linz’s et al. study concentrates on the affect of aggressive and violent exposure in subjects’ desensitization of perception of violence towards women. In my opinion it is hard to compare and relate two such studies because they are treating two different phenomena: exposure to television in general (i.e. medium exposure) vs. exposure to aggressive and violent material. Obviously, a treatment of reducing media exposure in itself contains reduction of violent and aggressive media behaviors, considering that 3rd and 4th graders were exposed to such material in the first place. This perhaps is another bad assumption on part of Robinson’s et al. study. If the subjects were exposed to aggressive and violent media, then, reducing such exposure might explain the reduction in aggressive and violent behavior in the treatment groups. If however there was no significant exposure to aggressive and violent media to being with, any perceived reduction after the treatment would have revealed that the preexisting aggressive behavior was due to other factors than media exposure. For this reason, it would have been beneficial if there were separate measures for the media literacy schooling from the turn-off period.

In conjunction with the above argument, another drawback of the Robinson’s et al. study is that does not account for the subject’s activities that have resulted from reduction in media viewing. It is possible that the media literacy by itself was good enough to enable the students to shift their attention from TV, videotape and videogame use, to reading books and other extracurricular activities. Thus, the same resulting effects of reduced aggressive behaviors might have been explained not by reduction of TV viewing, but by the increase in other beneficial activities such as reading books, outdoor activities, participation in extracurricular activities, community services, etc. This is an important aspect to be considered because Robinson et al. suggest that this sort of findings could inform public policy. If so, it is possible that some combination of reduction of media exposure in conjunction with extensive media literacy and encouragement of other social activities might yield better results.

In trying to compare and contrast Robinson’s et al. article with Centerwall’s, it appears that they are opposites of each other in a sense that Robinson’s treatment is reducing the exposure to media (up to a point of media turn-off), while the Centerwall article considers a ‘natural’ treatment of media exposure to a country’s population previously not exposed to TV. Both articles find media effects on aggressive behaviors according to their hypothesis, decreased and increased aggressive and violent behaviors respectfully. However, from the perspective of methodology and ability to claim the results as significant, one needs to be more cautious with the Centerwall’s article considering the very narrow definition of the design and limited variables used in the study. Though the results seem to be statistically significant, the introduction of other variables might have revealed moderating and/or controlling variables, or other independent variables actually responsible for the increase in homicide rates in the South African white population. One could as well argue that the distinction of the content (aggressive vs. non-aggressive) is important in the treatment where media exposure is increased or introduced anew, in contrast with the treatment where media exposure is reduced. Certainly, this line of thought assumed that there is distinction between exposure to aggressive media and exposure to non-aggressive media. Thus, the need to make this distinction might not be necessary in the case such as Robinson’s treatment because the mere act of reduction and restriction of media use reduces exposure to aggressive media. However, as I have argued above, I believe that content is important and I would attempt to distinguish between aggressive and violent media programming and non-violent and educational media programming. If for example the South African apartheid government had introduced the TV with educational and non-violent programming only, Centerwall’s study might have produced different results. From another point of view, one could argue that the actual increase in the homicide rates were due to the increased desperation and economical divide that TV only brought it to people’s attention as a moderating variable while the independent variable was the socio-economic and political condition in the country.

Conclusion

In conclusion I would like to say that these articles show that television, videotapes, and videogames exhibit properties that are labeled as media effects. However, I’m very cautious with such findings and would suggest a distinction between media originated effects vs. media mediated (or pass through) effects. Perhaps the studies would have presented more credibility if they had considered rival explanations for the results they obtained. For example, aggressive and violent media programming (especially the violent video games) could be considered media originated because most have no parallel in real life. On the other side, there are non-violent and non-aggressive media effects that could induce violent and aggressive behaviors on a segment of a population due to socio-economic and political programming. [194:631]

References:

Linz, D., Donnerstein, E., & Penrod, S. (1984). The Effects of Multiple Exposures to Filmed Violence Against Women. Journal of Communication, 34, 130-147

Feshbach, S. (1969). The Catharsis Effect: Research and Another View. In Baker. R.K. and Ball, S.J. (Eds.), Mass Media and Violence (461-473). Washington, D.C.: U.S. Government Printing Office

Centerwall, B.S. (1989). Exposure to television as a risk factor for violence. American Journal of Epidemiology, 129, 643-652

Robinson, T.N., Wilde, M.L., Navracruz, L.C., Haydel, K.F., & Varady, A. (2001). Effects of reducing children’s television and video game use on aggressive behavior: A randomized controlled trial. Archives of Pediatric Adolescent Medicine, 155, 17-23